Re: 2 MIME questions re: message/rfc822

Keith Moore writes:
> > > yeah, maybe I misunderstood, but my take was that the problem was 
> > > caused by the some intermediary which did not treat the 
> > > message/rfc822 body part as an opaque object.
> > If the body part were meant to be treated as an opaque object, it 
> > should have been labelled application/octet-stream.
> no.  you shouldn't be expected to hide useful information for the 
> recipient just so that the mail system will be transparent.
Agreed. But IMO, the list processor software is a recipient as far as 
the internet mail system is concerned. (You get a final delivery DSN 
when the message is delivered to the mailing list processor.)
> there are plenty of reasons for looking at a message to decide whether 
> to bounce or filter it.  there are far fewer valid reasons for 
> altering a message.
Converting to internal format and then back can transform invalid input 
into valid output. Is that bad?
(Wasn't there a similar thread years ago about lone CRs/LFs transformed 
into CRLF?)
Arnt